Development of a Semi-Automatic Segmentation Method for Retinal OCT Images Tested in Patients with Diabetic Macular Edema

• DOI: 10.1371/journal.pone.0082922
• Publication Date: 2013-12-26T21:30:54Z
• Authors: Yijun Huang; Ronald P. Danis; Jeong W. Pak; Shiyu Luo; James White; Xian Zhang; Ashwini Narkar; Amitha Domalpally
• Keywords: Diabetic macular edema; External limiting membrane

Purpose To develop EdgeSelect, a semi-automatic method for the segmentation of retinal layers in spectral domain optical coherence tomography images, and to compare results with manual method. Methods SD-OCT (Heidelberg Spectralis) scans 28 eyes (24 patients diabetic macular edema 4 normal subjects) were imported into customized MATLAB application, manually segmented by three graders at corresponding inner limiting membrane (ILM), segment/ellipsoid interface (ISe), retinal/retinal pigment epithelium (RPE), Bruch's (BM). The then independently same using semi-automated allowing guide/correct layer interactively. inter-grader reproducibility agreement locating positions between EdgeSelect methods assessed compared Wilcoxon signed rank test. Results varied from 0.15 1.21 µm, smaller than those (3.36–6.43 µm). test indicated had significantly better ranged 0.08 1.32 µm. There small differences two ILM (p = 0.012) BM (p<0.001), but these statistically indistinguishable ISe 0.896) RPE 0.771). Conclusions resulted good set subjects disease, suggesting that this approach is feasible OCT image analysis clinical trials.